There is considerable controversy surrounding current models of the structure of channels formed by the connexin gene family. A recent model of the Cx32 channel (Fleishman et al. 2004) that is based on the structure of Cx43, obtained by image processing of frozen hydrate 2D crystals (Linger et al. 1999) uses the third transmembrane segment, M3, to form the majority of the aqueous channel pore. This view is supported by results of SCAM (substituted cysteine accessibility method) studies of Cx32 intercellular channels (Skerrett et al., 2002) but not by the SCAM studies reported by Zhou et al. (1997) and Kronengold et al. (2003). These authors indicate that M1 and a portion of the first extracellular loop E1 form the pore of Cx32*43E1 and Cx46 functional hemichannels. We propose to use disulphide-trapping methods to test the helical contact points predicted by these disparate models. Our preliminary studies strongly support the view that the pore of connexin channels is formed primarily by M1 and demonstrate that the Cx32*43E1 hemichannel can be locked in a state dependent conformation by the formation of Cd2+-bridges between substituted cysteine residues in adjacent M1/E1 helices. Our results suggest, that the closure of connexin channels by loop-gating results from a rotation of the M1/E1 segment. We propose to continue studies of disulphide bond formation between substituted cysteines in the M1/E1 region to determine the proximity relations of residues located deeper in the hemichannel pore and to establish their functional correlates. The ability to lock channel channels in open and closed conformations provides a means to explore the nature of conformational changes that underlie voltage gating. We propose to use state-dependent lock to establish the relation between Vj and loop-gating, two forms of voltage gating that are common to all connexins. We will continue to use NMR to solve the structure of wild type and mutant N-termini. Our past studies have suggested that the structure of N-terminus is determined largely by hydrophobic interactions among conserved non-polar residues and by the presence of highly flexible turn in the vicinity of the 12th residue. We propose solve the structure of mutant peptides to test these hypotheses.